The general role in estimating prevalence is dividing the outcome of interest or an event (old or new) over the population group of interest at a specific period of time. Therefore, the sample size shoud be calculated out of the particular population of interest.